Lead Intelligence

Agent-based lead intelligence pipeline that finds, scores, and reaches high-value contacts through social graph analysis and warm path discovery.

When to Activate

• User wants to find leads or prospects in a specific vertical
• Building an outreach list for partnerships, sales, or fundraising
• Researching who to contact and the best path to them
• User mentions "find leads," "outreach list," "who should I contact," "warm intro"
• Need to score or rank a contact list by relevance
• Want to map common connections to discover warm intro paths

Tool Requirements

Required

• Exa MCP — Deep web search for people, companies, and signals (web_search_exa)
• X API — Follower/following graph, mutual connection analysis, recent activity (X_BEARER_TOKEN, plus write context credentials like X_CONSUMER_KEY, X_CONSUMER_SECRET, X_ACCESS_TOKEN, X_ACCESS_TOKEN_SECRET)

Optional (Enhance Results)

• LinkedIn — Use direct API if available, otherwise browser control for search, profile viewing, and message drafting
• Apollo/Clay API — For enrichment cross-referencing if user has access
• GitHub MCP — For developer-focused lead qualification
• Apple Mail / Mail.app — Draft cold or warm emails, but do not send automatically
• Browser Control — For LinkedIn and X when API coverage is insufficient or rate-limited

Pipeline Overview

┌─────────────┐     ┌──────────────┐     ┌─────────────────┐     ┌──────────────┐     ┌─────────────────┐
│ 1. Signal   │────>│ 2. Mutual    │────>│ 3. Discover     │────>│ 4. Enrich    │────>│ 5. Draft        │
│    Scoring  │     │    Ranking   │     │    Warm Paths   │     │              │     │    Outreach     │
└─────────────┘     └──────────────┘     └─────────────────┘     └──────────────┘     └─────────────────┘

Voice Before Outreach

Do not draft outreach messages from generic sales copy.

When the user's voice matters, first run brand-voice. Reuse its VOICE PROFILE throughout this skill instead of ad-hoc style redrafting.

If real-time X access is available, pull recent original posts before drafting. If unavailable, use provided samples or best-available repo/website material.

Stage 1: Signal Scoring

Search for high-signal people in the target vertical. Assign weights to each person based on:

Signal	Weight	Source
Role / title match	30%	Exa, LinkedIn
Industry match	25%	Exa company search
Recent relevant topic activity	20%	X API search, Exa
Follower count / influence	10%	X API
Geographic proximity	10%	Exa, LinkedIn
Engagement with your content	5%	X API interactions

Signal Search Method

# Step 1: Define target parameters
target_verticals = ["prediction markets", "AI tooling", "developer tools"]
target_roles = ["founder", "CEO", "CTO", "VP Engineering", "investor", "partner"]
target_locations = ["San Francisco", "New York", "London", "remote"]

# Step 2: Exa deep search for people
for vertical in target_verticals:
    results = web_search_exa(
        query=f"{vertical} {role} founder CEO",
        category="company",
        numResults=20
    )
    # Score each result

# Step 3: X API search for active voices
x_search = search_recent_tweets(
    query="prediction markets OR AI tooling OR developer tools",
    max_results=100
)
# Extract and score unique authors

Stage 2: Mutual Connection Ranking

For each scored target, analyze the user's social graph to find the warmest paths.

Ranking Model

1. Pull user's X following list and LinkedIn connections
2. For each high-signal target, check for shared connections
3. Apply social-graph-ranker model to score bridge value
4. Rank shared connections by:

Factor	Weight
Number of connections to target	40% — Highest weight; most connections = highest rank
Shared connection's current role / company	20% — Decision-maker vs. individual contributor
Shared connection's geography	15% — Same city = easier intro
Industry match	15% — Same vertical = natural intro
Shared connection's X account / LinkedIn	10% — Identifiability for outreach

Normalization rule:

Use social-graph-ranker when the user needs graph math itself, bridge ranking as a standalone report,
or explicit decay model tuning.

Within this skill, use the same weighted bridge model:

B(m) = Σ_{t ∈ T} w(t) · λ^(d(m,t) - 1)
R(m) = B_ext(m) · (1 + β · engagement(m))

Interpretation:

• Tier 1: High R(m) and direct bridge path → Request warm intro
• Tier 2: Medium R(m) and one-hop bridge path → Conditionally request intro
• Tier 3: No viable bridge → Use same lead record for direct cold outreach

Output Format

If user explicitly asks to split the ranking engine separately, visualize the math, or score the network 
outside of a full lead workflow, run `social-graph-ranker` independently first, then feed results back 
into this flow.

Mutual Ranking Report
=====================

#1  @mutual_handle (Score: 92)
    Name: Jane Smith
    Role: Partner @ Acme Ventures
    Location: San Francisco
    Connections to Target: 7
    Shared with: @target1, @target2, @target3, @target4, @target5, @target6, @target7
    Best Intro Path: Jane invested in Target1's company

#2  @mutual_handle2 (Score: 85)
    ...

Stage 3: Warm Path Discovery

For each target, find the shortest intro chain:

You ──[follows]──> Mutual A ──[invested in]──> Target Company
You ──[follows]──> Mutual B ──[co-founded]──> Target Person
You ──[attended]──> Event ──[also attended]──> Target Person

Path Types (Ordered by Warmth)

1. Direct mutual connection — You both follow / know the same person
2. Portfolio connection — Shared connection invested in or is an advisor to target company
3. Colleague / alumnus — Shared connection worked at same company or attended same school
4. Event overlap — Both attended same conference / initiative
5. Content interaction — Target engaged with shared connection's content and vice versa

Stage 4: Enrichment

For each qualified lead, pull:

• Full name, current title, company
• Company size, funding stage, recent news
• Recent X posts (last 30 days) — topics, tone, interests
• Shared interests with user (shared follows, similar content)
• Recent company events (product launch, funding round, hiring)

Enrichment Sources

• Exa: Company data, news, blog articles
• X API: Recent tweets, bio, follower count
• GitHub: Open source contributions (for developer-focused leads)
• LinkedIn (via browser): Full profile, experience, education

Stage 5: Outreach Draft

Generate personalized outreach messages for each lead. Drafts should align with the voice profile matched to source and the target channel.

Channel Rules

Email

• Use for highest-value cold outreach, warm intros, investor outreach, and partnership requests
• Default to drafting in Apple Mail / Mail.app when local desktop control is available
• Create drafts first; do not auto-send unless user explicitly requests
• Subject line should be concise and specific, not clever

LinkedIn

• Use when target is active on LinkedIn, mutual graph context is stronger on LinkedIn, or email confidence is low
• Prioritize API access if available
• Otherwise use browser control to view profile, recent activity, and draft messages
• Keep shorter than email; avoid false professional enthusiasm

X

• Use for high-context operator, builder, or investor outreach where public posting behavior matters
• Prioritize API access for search, timeline, and interaction analysis
• Fall back to browser control if needed
• DMs and public replies should be more compact than email and reference real content on target's timeline

Channel Selection Heuristic

Choose one primary channel in this order:

1. Email via warm intro
2. Direct email
3. LinkedIn DM
4. X DM or reply

Only use multiple channels if there is strong justification and pace does not look like spam.

Warm Intro Request (to shared connection)

Goals:

• One clear ask
• One concrete reason why this intro makes sense
• Provide an easy-to-forward bio if helpful

Avoid:

• Over-explaining your company
• Stacking social proof
• Sounding like a fundraising template

Direct Cold Outreach (to target)

Goals:

• Start with something specific and recent
• Explain why the fit is real
• Make a low-friction ask

Avoid:

• Generic praise
• Feature dumping
• Broad asks like "would love to connect"
• Forced rhetorical questions

Execution Mode

For each target, generate:

1. Recommended channel
2. Why that channel is best
3. Message draft
4. Optional follow-up draft
5. If email is the selected channel and Apple Mail is available, create a draft instead of just returning text

If browser control is available:

• LinkedIn: View target profile, recent activity, and mutual connection context, then draft or prepare message
• X: View recent posts or replies, then draft DM or public reply language

If desktop automation is available:

• Apple Mail: Create a draft email with subject, body, and recipient

Do not auto-send messages without explicit user approval.

Anti-Patterns

• Generic templates with no personalization
• Long paragraphs explaining your entire company
• Multiple asks in a single message
• False familiarity without specific details
• Bulk-sent messages with visible merge fields
• Reusing identical copy across email, LinkedIn, and X
• Platform-speak instead of authentic author voice

Configuration

Users should set the following environment variables:

# Required
export X_BEARER_TOKEN="..."
export X_ACCESS_TOKEN="..."
export X_ACCESS_TOKEN_SECRET="..."
export X_CONSUMER_KEY="..."
export X_CONSUMER_SECRET="..."
export EXA_API_KEY="..."

# Optional
export LINKEDIN_COOKIE="..." # For browser-use LinkedIn access
export APOLLO_API_KEY="..."  # For Apollo enrichment

Agents

This skill includes specialized agents in the agents/ subdirectory:

• signal-scorer — Search and rank prospects by relevance signals
• mutual-mapper — Map social graph connections and discover warm paths
• enrichment-agent — Pull detailed personal and company profiles
• outreach-drafter — Generate personalized messages

Usage Example

User: Help me identify the top 20 people in prediction markets I should contact

Agent Workflow:
1. signal-scorer searches Exa and X for prediction market leaders
2. mutual-mapper checks user's X social graph for shared connections
3. enrichment-agent pulls company data and recent activity
4. outreach-drafter generates personalized messages for top-ranked prospects

Output: Ranked list with warm paths, voice-profile summary, and in-app drafts for specific channels

Related Skills

• brand-voice for normative voice capture
• connections-optimizer for pre-outreach network pruning and expansion